KDirModelV2, KDirListerV2 and UDSEntryV2 suggestions

[Mark]

On Tue, Feb 5, 2013 at 9:01 AM, Mark <markg85 at gmail.com> wrote:
> On Tue, Feb 5, 2013 at 8:03 AM, Frank Reininghaus
> <frank78ac at googlemail.com> wrote:
>> Hi Mark,
>>
>> 2013/2/4 Mark:
>>> I really really really dislike KDirModel and friends (KDirLister,
>>> KFileItem).
>>
>> before you get even more emotional in your next reply to this thread,
>> please consider reading
>>
>> http://www.kde.org/code-of-conduct/
>>
>> I sort of got used to reading disrespectful messages with little
>> useful content from a small minority of our user base, but I feel at
>> least mildly offended when I read such statements on a developer-only
>> mailing list. Not only because the classes you mention have been
>> serving us well for quite some time and David does an awesome job
>> maintaining them (he fixes a bug in no time at all as soon as I throw
>> a half-ready unit test at him), but also because they contain some
>> contributions from myself. Yes, it's mostly unit tests and one-line
>> patches, but in some cases, I needed a few hours of backtrace and code
>> reading before I could figure out what's going wrong [1].
>>
>> Thanks for your understanding and best regards,
>> Frank
>>
>> [1] https://bugs.kde.org/show_bug.cgi?id=196695
>
> Hi Frank,
>
> the "really really really dislike" part was meant to be taken
> seriously. Specially not after adding a screenshot with it's memory
> usage. I like diving into those classes and try to figure out more
> optimal ways of doing things, i certainly don't hate them nor
> disrespect the authors of them.
>
> The thing i'm puzzling most with right now is how i can optimize
> UDSEntry. Internally it's a hash and that very visible in profiling.
> Also in KFileItem one part that i find a little strange is this line:
> http://quickgit.kde.org/?p=kdelibs.git&a=blob&h=6667a90ee9e1d57488bb7e085167658f2fb9f172&hb=533b48c610319f3ad67e6f5f0cbb65028b009b8f&f=kio%2Fkio%2Fkfileitem.cpp
> (line 290). That line is causing a chain of performance penalties.
> Which is very odd because i'm testing this benchmark with 500.000
> files, not directories. It should not even end up in that if.
>
> Or am i reading massif wrong..? Massif shows me that line for KUrl
> data consumption.. One of the highest memory consumptions.

<_< typo.. "part was meant to be taken" should be "part was't meant to
be taken" Sorry for that.